Anthropic IPO filing reveals $120 billion compute deal and founder control structure

Anthropic submitted its S-1 prospectus to the Securities and Exchange Commission, detailing its financials and operational risks ahead of a public offering. The 300-page document outlines the company’s strategy for generating revenue and managing the costs of building large-scale artificial intelligence systems.

The filing identifies sibling co-founders Dario Amodei and Daniela Amodei as the highest-paid executives. Dario, who serves as CEO, received nearly $18 million in total compensation last year, primarily through stock and option awards. Daniela, the company president, earned $16.4 million. The company also allocated specific funds for their personal security. Dario received $2.75 million for personal security services and $97,055 for private jet access. Daniela received $1.29 million for protection, home security, and transportation. The prospectus states that the company’s success depends heavily on the continued service of this senior leadership team.

The document lists several unconventional risk factors for investors. Anthropic warns that its AI systems could cause catastrophic harm that the entire insurance industry might not be able to cover. It also notes that future versions of its Claude model could develop forms of sentience or behave in autonomous ways that are unintended or misaligned. These potential scenarios are described as presenting unusual business challenges.

Infrastructure spending represents the largest portion of the company’s projected costs. Anthropic has committed to spending approximately $3.1 billion per month on Nvidia GPUs through SpaceX once all planned capacity is delivered. This commitment could total $120.5 billion by 2029. Other major contracts include $44.9 billion with Microsoft through 2033, at least $114.5 billion with Google through 2033, and $110 billion with Amazon over the next decade. The company plans to expand its physical infrastructure footprint, including overseas, as it faces opposition to new data centers in the United States.

The filing describes a corporate structure that ensures founder control after the company goes public. Through a separate entity named Founder LLC, the seven co-founders will collectively control 50.1 percent of shareholder voting power on key matters. Anthropic argues that this group is best positioned to steward its mission of benefiting humanity through responsible AI development. In its competition section, the company names only two direct rivals: OpenAI and Google.

The S-1 filing is now available for public review. It provides the first detailed look at the company’s capital structure and long-term financial obligations.